F9 — Ask AI
Ask anything without leaving your window.
Hold F9 and ask in plain speech, or press Ctrl+F9 to open the panel and start a screenshot question. Choose a screen or app window; an image-capable model analyzes the image. Answers stream into the side overlay, and chats with screenshot attachments are saved locally.
F9 → “summarize this thread in 3 bullets”
1. Ship date moved to Friday
2. Ana owns the changelog
3. You own the release note
How F9 works
- 1
Hold F9 and ask
Ask anything — explanations, code help, writing drafts, quick facts — without switching tabs.
- 2
Answer streams beside you
The overlay renders markdown, tables, and math so answers are readable, not a wall of text.
- 3
Follow up in context
Enable continuous mode to keep the thread across questions. Chats are saved locally.
- 4
Ask about a screenshot
Press Ctrl+F9, choose Screenshot, select a screen or app window, and ask your question. The image goes with the F9 request to its assigned provider; use a vision-capable model.
- 5
Copy or dismiss
Copy what you need into your app, or dismiss and keep working. Your cursor never moves.
Why Ask AI saves time
- No tab-switching: answers appear beside your work
- Markdown + tables + math rendering built in
- Continuous mode for multi-turn questions
- Optional F9 screenshot questions for a selected screen or app window; image-capable model required
- Chat history and screenshot attachments stored locally, wipeable anytime
- Cloud providers use your own key; local Ollama is available without a cloud provider key
F9 FAQ
In a side overlay next to your current window. Your cursor and app stay put — copy out only what you need.
Yes. Turn on continuous mode and Flunkey keeps the conversation context across follow-ups.
Yes, locally on your machine with markdown, tables, and math rendering. Wipe everything in Settings → Wipe all data.
Press Ctrl+F9 to open the panel, choose Screenshot, select a screen or app window, then ask. Only F9 sends screenshots; they are attached to the locally saved chat. Choose a vision-capable model for image analysis.
Assign a provider and model to F9 in Models & Providers. Cloud providers include Groq, OpenAI, Gemini, and OpenRouter; Ollama can run local models without a cloud key. Screenshot analysis requires a vision-capable model.
Yes. Ask for explanations, regex, SQL, or error help and paste the result into your editor.
Cloud requests go directly to the provider assigned to F9. If you assign Ollama, requests go to your configured Ollama server, normally localhost. Flunkey does not proxy the request.
Canonical: https://www.flunkey.space/features/ask-ai · Also reachable via /f9
Compare: Compare Flunkey vs Superwhisper · See who dictates anywhere with Flunkey voice typing.